Metadata expansion method based on semantic analysis

Abstract

The invention relates to the technical field of metadata expansion, and discloses a semantic analysis-based metadata expansion method, which comprises the following steps of: inputting a search text by a target user; extracting keywords according to the search text, and determining the keywords as search keywords; acquiring history browsing course data of the target user in the target database, and performing pushing professional judgment according to the history browsing course data and the search keyword; pushing courses related to the pushing specialty; obtaining current browsing data; and adjusting course pushing according to the current browsing data. According to the metadata extension method based on semantic analysis, the professional tendency of a target user to a course which the target user wants to browse can be judged according to historical browsing course data or a current-day webpage browsing record or current browsing pushing course fragment data, and the course is pushed to the target user in a targeted manner; in the course browsing process, the course pushing difficulty degree and the course related knowledge points can be adjusted according to the current browsing data, and the user requirements are met.

Description

Technical Field

[0001] The present invention relates to the technical field of metadata extension, and in particular to a metadata extension method based on semantic analysis. Background Art

[0002] With the continuous increase in the amount of information on the Internet, how to effectively find and discover information that is useful to users has become a key issue that information systems need to solve. At present, information retrieval technology can be divided into three categories: full-text retrieval, data retrieval and semantic retrieval. Among them, semantic retrieval emphasizes knowledge-based and semantic matching, and has good application prospects in information retrieval, especially in knowledge-based information retrieval.

[0003] In the prior art, in some course learning systems, when a user wants to search for the course data to be studied, the user is generally required to enter a search text, the system extracts the keywords of the search text, and pushes corresponding courses to the user based on the keywords, which is not convenient for expansion based on the user's historical browsing data. In reality, courses of multiple majors may be related to each other, that is, multiple majors may have courses corresponding to keywords. The metadata expansion method of the prior art generally pushes courses of all majors corresponding to the keywords to the user, which is not convenient for targeted push based on the user's inclination towards the corresponding major. In addition, due to individual differences in course learning, each user has different levels of acceptance of the difficulty of the course and different levels of understanding and mastery of knowledge points. Therefore, the currently pushed courses may not necessarily meet the user's learning needs. If they are not adjusted in time, it will be difficult to meet the actual needs of users. Summary of the Invention

[0004] The present invention provides a metadata expansion method based on semantic analysis, which can judge the user's professional browsing tendency based on historical browsing course data or daily web browsing records or currently browsing and pushing course fragment data, and push course content to the user in a targeted manner. It can adjust the course push according to the current browsing data so that the difficulty level of the pushed course or the relevant knowledge points can be adapted by the user. It solves the problem mentioned in the above background technology in the metadata expansion method in the prior art that courses are pushed to users only based on keywords, which cannot meet the needs of users who prefer a certain profession, and due to individual differences, each user has different acceptance levels of the difficulty level of the course and different understanding and mastery of the knowledge points. If timely adjustments are not made, it is difficult to meet the user's learning needs.

[0005] The present invention provides the following technical solution: a metadata expansion method based on semantic analysis, comprising:

[0006] The target user enters the search text;

[0007] Extract keywords based on the search text and define them as search keywords;

[0008] Obtain the target user's historical course browsing data in the target database, and push professional judgment based on the historical course browsing data and search keywords;

[0009] Push and push professional related courses;

[0010] Get current browsing data;

[0011] Adjust course push based on current browsing data.

[0012] As an optional solution to the metadata expansion method based on semantic analysis of the present invention, the method of obtaining the target user's historical course browsing data in the target database and pushing professional judgment based on the historical course browsing data and search keywords includes:

[0013] Obtain the target user's historical course browsing data in the target database;

[0014] If there is historical course browsing data, the major judgment will be pushed based on the historical course browsing data;

[0015] If there is no historical course browsing data, the course clips corresponding to the search keywords will be pushed, and the professional judgment will be pushed based on the data of the target users browsing the course clips.

[0016] As an optional solution of the metadata expansion method based on semantic analysis of the present invention, the pushing of professional judgment based on historical browsing course data includes:

[0017] Extract all keywords corresponding to historical browsing course data;

[0018] If the search keyword is included, the target user's historical course browsing data corresponding to the search keyword is obtained. The historical course browsing data includes the major of the historical course browsing, the total duration of the historical course browsing, and the target user's historical course browsing duration. The total duration of the historical course browsing and the target user's historical course browsing duration are recorded as t1 and t2 respectively.

[0019] Calculate the ratio of the target user's history course browsing time t2 to the total history course browsing time t1 for each major, denoted as A. Ratio A = history course browsing time t2 ÷ total history course browsing time t1;

[0020] Sort the ratio A corresponding to each major from large to small;

[0021] The major with the largest ratio A is determined to be the recommended major;

[0022] If the search keyword is not included, the target user's web browsing history is obtained;

[0023] Extract all web page records browsed by the target user on the day the target user enters the search text;

[0024] Extract all majors corresponding to each web page record;

[0025] Get the common majors in each web page record and define them as push majors.

[0026] As an optional solution to the metadata expansion method based on semantic analysis of the present invention, the method of pushing the course segments corresponding to the search keywords and making a push professional determination based on the data of the target user browsing the course segments includes:

[0027] Obtain all courses in the target database that correspond to the search keyword;

[0028] A course segment was extracted from the courses corresponding to each major;

[0029] Push course clips to target users;

[0030] Obtain the target user's browsing course segment data, which includes the course segment's major, the total duration of the course segment, and the target user's browsing time. The total duration of the course segment and the target user's browsing time are recorded as t3 and t4, respectively.

[0031] Calculate the ratio of the target user's course segment browsing time t4 to the total course segment duration t3 for each major, denoted as B. Ratio B = course segment browsing time t4 ÷ total course segment duration t3;

[0032] Sort the ratio B corresponding to each major from large to small;

[0033] The major with the largest ratio B is determined to be the push major.

[0034] As an optional solution to the metadata expansion method based on semantic analysis of the present invention, the push of courses related to the pushed major includes:

[0035] Set the total number of courses to be pushed, recorded as N;

[0036] Obtain the courses that correspond to the search keyword and the recommended major in the target database and define them as keyword courses;

[0037] Push keyword courses to target users based on the set total number of course pushes N.

[0038] As an optional solution of the metadata expansion method based on semantic analysis of the present invention, the adjustment of course push according to the current browsing data includes:

[0039] Get the time the target user spends watching the keyword course and the total time the keyword course lasts, recorded as t5 and t6 respectively;

[0040] Calculate the ratio of the target user's viewing time of the keyword course to the total duration of the keyword course, recorded as C, ratio C = viewing time t5 of the keyword course ÷ total duration t6 of the keyword course;

[0041] If the ratio C is ≥ 90%, the target user's knowledge mastery is judged, and course push adjustments are made based on the target user's knowledge mastery.

[0042] If the ratio C is less than 90%, the difficulty level of the course will be adjusted.

[0043] As an optional solution to the metadata expansion method based on semantic analysis described in the present invention, the method of determining the target user's knowledge mastery and adjusting the course push based on the target user's knowledge mastery includes:

[0044] Obtain the pause, fast-forward, and rewind status of target users during the course viewing process;

[0045] If there is a pause or fast-forward, the course difficulty level will be adjusted;

[0046] If there is no pause or fast forward, but there is rewind, the knowledge point push adjustment will be made;

[0047] If there is no pause, fast forward, or rewind, no adjustment will be made.

[0048] As an optional solution to the metadata expansion method based on semantic analysis of the present invention, the course difficulty level push adjustment includes:

[0049] Classify keyword courses according to the difficulty level of the courses, where the keyword course classification includes difficult and easy;

[0050] Get the category of the keyword course that the target user is currently watching;

[0051] If the category of the keyword course currently being viewed is easy, adjust the category of the keyword course to difficult;

[0052] If the classification of the keyword course currently being watched is difficult, a questionnaire will be pushed to the target user, asking whether the keyword course currently being watched by the target user is easy;

[0053] If the target user answers simply, then ask the target user whether they understand the knowledge points involved in the keyword course they are currently watching;

[0054] If the target user answers that they understand, no adjustment will be made;

[0055] If the target user answers that they do not understand, all knowledge points in the keyword course currently watched by the target user are obtained, and courses corresponding to other knowledge points under the major that do not include all knowledge points in the keyword course currently watched by the target user are pushed to the target user;

[0056] If the target users have difficulty answering, adjust the classification of the keyword course to easy.

[0057] As an optional solution of the metadata expansion method based on semantic analysis of the present invention, the knowledge point push adjustment includes:

[0058] Get the time point before and after rewinding the target user's current course, recorded as P1 and P2 respectively;

[0059] Extract the knowledge points between time point P2 and time point P1;

[0060] Refine and organize the extracted knowledge points to form knowledge point keywords;

[0061] Obtain the courses that correspond to the search keywords, knowledge point keywords, and recommended majors in the target database and define them as knowledge point courses;

[0062] Get the duration of the knowledge point segment corresponding to the knowledge point keyword in the course currently watched by the target user, recorded as t7;

[0063] Extract all knowledge points in the target user's current course;

[0064] Get the duration of each knowledge point segment corresponding to the target user's current course, and record them as M1, M2, ..., Mn respectively;

[0065] Calculate the sum of the duration of the corresponding knowledge point segments of all knowledge points in the target user's current course, denoted as M, M = M1 + M2 + ... + Mn;

[0066] Calculate the ratio of the duration t7 of the knowledge point segment corresponding to the knowledge point keyword in the target user's current course to the sum of the durations M of the knowledge point segments corresponding to all knowledge points in the target user's current course, denoted as D. Ratio D = duration t7 ÷ sum of durations M;

[0067] According to the total number of course pushes N and the ratio D, calculate the number of knowledge point course pushes, recorded as N1, the number of knowledge point course pushes N1 = total number of pushes N × ratio D;

[0068] According to the total number of course pushes N and the number of knowledge point course pushes N1, calculate the number of keyword course pushes, recorded as N2, the number of keyword course pushes N2 = the total number of pushes N - the number of knowledge point course pushes N1;

[0069] Adjust the course push according to the number of knowledge point courses pushed N1 and the number of keyword courses pushed N2, and push knowledge point courses and keyword courses to target users.

[0070] The present invention has the following beneficial effects:

[0071] 1. This metadata expansion method based on semantic analysis can more accurately judge the target user's professional inclination towards the courses they want to browse based on historical course browsing data, daily web browsing records, or currently browsed course fragment data, thereby determining the pushed major and pushing courses to the target user in a targeted manner, so that the final pushed courses better meet the target user's browsing and learning needs and improve the target user's learning efficiency.

[0072] 2. This metadata expansion method based on semantic analysis can adjust the difficulty level of course push and relevant knowledge points of the course according to the current browsing data during the course browsing process, so that the difficulty level of subsequent pushed courses can be adapted to the target users, and the relevant knowledge points in the subsequent pushed courses can be easily understood and mastered by the target users, meeting the users' actual learning needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 This is a block diagram of the metadata expansion method based on semantic analysis of the present invention. DETAILED DESCRIPTION

[0074] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0075] Embodiment 1, a metadata expansion method based on semantic analysis, includes:

[0076] The target user enters the search text. Specifically, the search text can be a word, a sentence, or a paragraph;

[0077] Extract keywords from the search text and define them as search keywords. For example, the search keywords extracted from the search text may be "circuit board";

[0078] Obtain the target user's historical course browsing data in the target database, and push professional judgment based on the historical course browsing data and search keywords;

[0079] Push and push professional related courses;

[0080] Get current browsing data;

[0081] Adjust course push based on current browsing data.

[0082] Through the metadata expansion method based on semantic analysis, it is possible to more accurately judge the professional inclination of target users towards the courses they want to browse based on historical course browsing data, the day's web browsing records, or the currently browsed pushed course fragment data, thereby determining the pushed major and pushing courses to target users in a targeted manner based on the pushed major, so that the pushed courses better meet the browsing and learning needs of target users and improve the learning efficiency of target users; in the process of browsing courses, the difficulty level of course push and the relevant knowledge points of the course can be adjusted based on the current browsing data, the difficulty level of the pushed courses can be adapted to the target users, and the push of relevant knowledge points that are easy for the target users to understand and master can meet the actual learning needs of users.

[0083] Example 2: This example is an improvement based on Example 1. Specifically, the target user's historical course browsing data in the target database is obtained, and the professional judgment is pushed based on the historical course browsing data and search keywords, including:

[0084] Obtain the target user's historical course browsing data in the target database;

[0085] If there is historical course browsing data, the major judgment will be pushed based on the historical course browsing data;

[0086] If there is no historical course browsing data, the course clips corresponding to the search keywords will be pushed, and the professional judgment will be pushed based on the data of the target users browsing the course clips.

[0087] Push professional judgment based on historical course browsing data, including:

[0088] Extract all keywords corresponding to historical browsing course data;

[0089] If the search keyword is included, the historical course browsing data of the target user corresponding to the search keyword is obtained. The historical course browsing data includes the major of the historical course browsing, the total duration of the historical course browsing and the historical course browsing time of the target user. The total duration of the historical course browsing and the historical course browsing time of the target user are recorded as t1 and t2 respectively. Specifically, the search keyword is included if there is a keyword that is the same as the search keyword "circuit board" among all the keywords extracted from the historical course browsing data; the total duration of the historical course browsing is the total duration of the courses corresponding to the search keyword and browsed by the target user under each major, and the target user's historical course browsing time is the total duration of the courses corresponding to the search keyword under each major. For example, if all keywords corresponding to the historical course browsing data are "circuit board", "wireless communication" and "signal processing", the search keyword "circuit board" is included. For another example, the courses corresponding to the search keyword "circuit board" browsed by the target user include "Electronic Information Engineering" and "Mechatronics". The target user browsed the "Electronic Information Engineering" course for 35 minutes, and the total duration of the "Electronic Information Engineering" course was 40 minutes. The target user browsed the "Mechatronics" course for 5 minutes, and the total duration of the "Mechatronics" course was 50 minutes.

[0090] Calculate the ratio of the target user's historical course browsing time t2 to the total historical course browsing time t1 for each major, denoted as A. Ratio A = historical course browsing time t2 ÷ total historical course browsing time t1, i.e., the ratio A for "Electronic Information Engineering Major" is A = 35 ÷ 40 = 0.875, and the ratio A for "Mechatronics Major" is A = 5 ÷ 50 = 0.1.

[0091] Sort the ratio A corresponding to each major from largest to smallest, i.e., the ratio A of "Electronic Information Engineering" is greater than the ratio A of "Mechatronics". By comparing the ratio A corresponding to each major, we can determine the target user's professional preference for browsing courses.

[0092] The major with the largest ratio A is determined to be the recommended major. That is, the ratio of "Electronic Information Engineering" is the largest, so "Electronic Information Engineering" is the recommended major.

[0093] If the search keyword is not included, the target user's web browsing history is obtained. Specifically, the search keyword is not included if there is no keyword that is the same as the search keyword among all the extracted keywords. For example, if all the keywords corresponding to the historical browsing course data are "wireless communication" and "signal processing", the search keyword "circuit board" is not included;

[0094] Extract all web page records browsed by the target user on the day the search text was entered, for example: the target user browsed four web pages on the day the search text was entered;

[0095] Extract all majors corresponding to each webpage record. For example, the first webpage has "Mechatronics" and "Electronic Information Engineering", the second webpage has "Electronic Information Engineering" and "Software Technology", the third webpage has "Communication Engineering" and "Electronic Information Engineering", and the third webpage has "Internet of Things Application Technology" and "Electronic Information Engineering";

[0096] The common majors in each web page record are obtained and determined as push majors. Specifically, by obtaining the common majors in each web page record, the target user's professional inclination towards browsing courses can be determined. For example, if the common major in four web pages is "Electronic Information Engineering", then "Electronic Information Engineering" will be positioned as the push major.

[0097] Push course clips corresponding to the search keywords, and make professional judgments based on the data of target users browsing course clips, including:

[0098] Get all the courses of the major corresponding to the search keyword in the target database, that is, get all the courses of the major corresponding to the search keyword "circuit board" in the target database. For example, the courses corresponding to the search keyword "circuit board" include "communication engineering major", "electronic information engineering major" and "mechatronics major".

[0099] A course segment is extracted from the courses corresponding to each major. The extracted course segment is a classic course segment with pictures. That is, a classic course segment with pictures is extracted from the courses corresponding to "Communication Engineering Major" and "Electronic Information Engineering Major".

[0100] Push course clips to target users. Specifically, push the extracted course clips for each major to target users.

[0101] Obtain target user browsing course segment data, the course segment data includes the major to which the course segment belongs, the total duration of the course segment and the target user's course segment browsing time. The total duration of the course segment and the target user's course segment browsing time are recorded as t3 and t4 respectively. Specifically, the total duration of the course segment is the total duration of the course segment under each major, and the target user's course segment browsing time is the target user's browsing time of the course segment under each major. For example, set the total duration t3 of the course segment of "Communication Engineering Major" to 5 minutes, and the course segment browsing time t4 to 1 minute; set the total duration t3 of the course segment of "Electronic Information Engineering Major" to 5 minutes, and the course segment browsing time t4 to 5 minutes;

[0102] Calculate the ratio of the target user's course segment browsing time t4 to the total course segment duration t3 for each major, denoted as B. Ratio B = course segment browsing time t4 ÷ total course segment duration t3. That is, the ratio B for "Communication Engineering" is 1 ÷ 5 = 0.2, and the ratio B for "Electronic Information Engineering" is 5 ÷ 5 = 1.

[0103] Sort the ratio B corresponding to each major from large to small. Specifically, by comparing the ratio B corresponding to each major, the target user's professional preference for browsing courses can be determined, that is, the ratio B of "Electronic Information Engineering Major" is greater than the ratio B of "Communication Engineering Major";

[0104] The major with the largest ratio B is determined to be the pushed major, that is, the ratio of "Electronic Information Engineering Major" is the largest, so "Electronic Information Engineering Major" is the pushed major.

[0105] Example 3: This example is an improvement made on the basis of Example 1. Specifically, the courses related to the pushed majors include:

[0106] Set the total number of courses pushed, recorded as N, for example: the total number of courses pushed N = 6;

[0107] Obtain the courses that correspond to the search keyword and the recommended major in the target database and define them as keyword courses;

[0108] According to the set total number of course pushes N, keyword courses are pushed to the target users, that is: 6 groups of keyword courses are pushed to the target users.

[0109] Example 4: This example is an improvement based on Example 1. Specifically, the course push is adjusted according to the current browsing data, including:

[0110] Obtain the time the target user spent watching the keyword course and the total time the keyword course lasts, which are recorded as t5 and t6 respectively. Specifically, determine whether the target user is watching the course by whether the target user's gaze is at the course display location, thereby obtaining the time the target user spent watching the keyword course;

[0111] Calculate the ratio of the target user's viewing time for the keyword course to the total duration of the keyword course, denoted as C. Ratio C = viewing time t5 of the keyword course ÷ total duration t6 of the keyword course. Specifically, the ratio C can be used to determine the target user's concentration level, and thus the target user's degree of adaptation to the current keyword course. Based on the target user's degree of adaptation, decide whether to adjust the course push and how to adjust it.

[0112] If the ratio C is ≥ 90%, the target user's knowledge mastery is judged, and course push adjustments are made based on the target user's knowledge mastery. Specifically, when the ratio C is ≥ 90%, it indicates that the target user is relatively adaptable to the current keyword course, thereby further judging the target user's knowledge mastery.

[0113] If the ratio C is less than 90%, the course difficulty level will be adjusted. Specifically, when the ratio C is less than 90%, it indicates that the target user is not very adaptable to the current keyword course, and the course difficulty level will be adjusted directly.

[0114] Determine the target user's knowledge level and adjust the course delivery based on the target user's knowledge level, including:

[0115] Obtain the pause, fast-forward, and rewind status of target users during the course viewing process;

[0116] If there is a pause or fast-forward, the course difficulty level will be adjusted. Specifically, if there is a pause or fast-forward, it indicates that the target user may not be very adaptable to the difficulty level of the course, so the course difficulty level will be adjusted.

[0117] If there is no pause or fast-forward, but there is rewind, the knowledge point push adjustment is made. Specifically, if there is no pause or fast-forward, but there is rewind, it indicates that the target users may not have fully mastered the knowledge points in this course, so the corresponding knowledge point push adjustment is made;

[0118] If there is no pause, fast forward, or rewind, no adjustment will be made. Specifically, if there is no pause, fast forward, or rewind, it indicates that the target user has completed the entire course in the order and speed of the course content, further indicating that the target user may be fully adapted to the difficulty level of the course and can fully grasp the knowledge points in the course, so the pushed course will not be adjusted and will be pushed according to the original settings.

[0119] Adjustments to the course difficulty level are pushed, including:

[0120] Classify keyword courses according to the difficulty level of the courses, where the keyword course classification includes difficult and easy;

[0121] Get the category of the keyword course that the target user is currently watching;

[0122] If the category of the keyword course currently being viewed is easy, adjust the category of the keyword course to difficult;

[0123] If the classification of the keyword course currently being watched is difficult, a questionnaire will be pushed to the target user, asking whether the keyword course currently being watched by the target user is easy;

[0124] If the target user answers simply, then ask the target user whether they understand the knowledge points involved in the keyword course they are currently watching;

[0125] If the target user answers that they understand, no adjustment will be made;

[0126] If the target user answers "I don't understand," all the knowledge points in the keyword course the target user is currently watching are obtained, and courses corresponding to other knowledge points under the major that do not include all the knowledge points in the keyword course the target user is currently watching are pushed to the target user. Specifically, if the target user answers "I don't understand," it may be that the target user does not understand the knowledge points involved in the keyword course they are currently watching. Therefore, other knowledge points can be pushed to the target user for learning;

[0127] If the target users have difficulty answering, adjust the classification of the keyword course to easy.

[0128] Adjustments to knowledge point push, including:

[0129] Get the time point before and after rewinding the target user's current course, recorded as P1 and P2 respectively;

[0130] Extract the knowledge points between time point P2 and time point P1;

[0131] Refine and organize the extracted knowledge points to form knowledge point keywords;

[0132] Obtain the courses that correspond to the search keywords, knowledge point keywords, and recommended majors in the target database and define them as knowledge point courses;

[0133] Get the duration of the knowledge point segment corresponding to the knowledge point keyword in the course currently watched by the target user, recorded as t7. For example, set the duration t7 of the knowledge point segment corresponding to the knowledge point keyword in the course currently watched by the target user to 4.8 minutes;

[0134] Extract all knowledge points in the target user's current course. For example, set the target user's current course to involve 4 knowledge points.

[0135] Get the duration of each knowledge point segment corresponding to the target user's current course, and record them as M1, M2, ..., Mn. For example, set the durations of the four knowledge point segments as: M1 = 6.5 min, M2 = 6 min, M3 = 4.8 min, M4 = 6.7 min, where M3 is the duration t7 of the knowledge point segment corresponding to the knowledge point keyword in the target user's current course;

[0136] Calculate the sum of the duration of all the knowledge point segments corresponding to the target user's current course, denoted as M, where M = M1 + M2 + ... + Mn, i.e., the sum of the duration M = 6.5 + 6 + 4.8 + 6.7 = 24. Specifically, in this process, the duration of some unimportant course segments without knowledge point explanations in the current course will be discarded, and only the duration of the course segments corresponding to all knowledge point explanations will be calculated;

[0137] Calculate the ratio of the duration t7 of the knowledge point segment corresponding to the knowledge point keyword in the target user's current course to the sum of the duration M of the knowledge point segments corresponding to all knowledge points in the target user's current course, denoted as D. Ratio D = duration t7 ÷ sum of durations M, i.e., ratio D = 4.8 ÷ 24 = 0.2. Specifically, calculate the ratio D by summing the duration of the course segment corresponding to the knowledge point keyword in this course and the duration of the course segment corresponding to all knowledge point explanations in this course. Discard the duration of some content in this course that does not have key points or knowledge points explained, making the calculated ratio D more accurate, thereby ensuring that subsequent push adjustments are more in line with the needs of the target user.

[0138] According to the total number of course pushes N and the ratio D, calculate the number of knowledge point course pushes, recorded as N1. The number of knowledge point course pushes N1 = the total number of pushes N × the ratio D, that is: the number of knowledge point course pushes N1 = 6 × 0.2 = 1.2;

[0139] According to the total number of course pushes N and the number of knowledge point course pushes N1, calculate the number of keyword course pushes, recorded as N2, the number of keyword course pushes N2 = total number of pushes N - number of knowledge point course pushes N1, that is: the number of keyword course pushes N2 = 6 - 1.2 = 4.8;

[0140] Adjust the course push according to the number of knowledge point courses pushed N1 and the number of keyword courses pushed N2, and push knowledge point courses and keyword courses to the target users. Specifically, since the push quantity is generally an integer, the push quantity of knowledge point courses and the push quantity of keyword courses are rounded up to the integer, that is: push 1 knowledge point course to the target user and push 5 keyword courses.

[0141] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0142] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A metadata expansion method based on semantic analysis, characterized by: include: The target user enters the search text; Extract keywords based on the search text and define them as search keywords; Obtain the target user's historical course browsing data in the target database, and push professional judgment based on the historical course browsing data and search keywords; Push and push professional related courses; Get current browsing data; Adjust course push based on current browsing data.

2. The metadata expansion method based on semantic analysis according to claim 1, characterized in that: The process of obtaining the target user's historical course browsing data in the target database and pushing a major decision based on the historical course browsing data and search keywords includes: Obtain the target user's historical course browsing data in the target database; If there is historical course browsing data, the major judgment will be pushed based on the historical course browsing data; If there is no historical course browsing data, the course clips corresponding to the search keywords will be pushed, and the professional judgment will be pushed based on the data of the target users browsing the course clips.

3. The metadata expansion method based on semantic analysis according to claim 2, characterized in that: The push major determination based on historical course browsing data includes: Extract all keywords corresponding to historical browsing course data; If the search keyword is included, the target user's historical course browsing data corresponding to the search keyword is obtained. The historical course browsing data includes the major of the historical course browsing, the total duration of the historical course browsing, and the target user's historical course browsing duration. The total duration of the historical course browsing and the target user's historical course browsing duration are recorded as t1 and t2 respectively. Calculate the ratio of the target user's history course browsing time t2 to the total history course browsing time t1 for each major, denoted as A. Ratio A = history course browsing time t2 ÷ total history course browsing time t1; Sort the ratio A corresponding to each major from large to small; The major with the largest ratio A is determined to be the recommended major; If the search keyword is not included, the target user's web browsing history is obtained; Extract all web page records browsed by the target user on the day the target user enters the search text; Extract all majors corresponding to each web page record; Get the common majors in each web page record and define them as push majors.

4. The metadata expansion method based on semantic analysis according to claim 3, characterized in that: The push of course segments corresponding to the search keyword and the determination of the push major based on the data of the target user browsing the course segments include: Obtain all courses in the target database that correspond to the search keyword; A course segment was extracted from the courses corresponding to each major; Push course clips to target users; Obtain the target user's browsing course segment data, which includes the course segment's major, the total duration of the course segment, and the target user's browsing time. The total duration of the course segment and the target user's browsing time are recorded as t3 and t4, respectively. Calculate the ratio of the target user's course segment browsing time t4 to the total course segment duration t3 for each major, denoted as B. Ratio B = course segment browsing time t4 ÷ total course segment duration t3; Sort the ratio B corresponding to each major from large to small; The major with the largest ratio B is determined to be the push major.

5. The metadata expansion method based on semantic analysis according to claim 4, characterized in that: The courses related to push and push majors include: Set the total number of courses to be pushed, recorded as N; Obtain the courses that correspond to the search keyword and the recommended major in the target database and define them as keyword courses; Push keyword courses to target users based on the set total number of course pushes N.

6. The metadata expansion method based on semantic analysis according to claim 5, characterized in that: The adjustment of course push based on current browsing data includes: Get the time the target user spends watching the keyword course and the total time the keyword course lasts, recorded as t5 and t6 respectively; Calculate the ratio of the target user's viewing time of the keyword course to the total duration of the keyword course, recorded as C, ratio C = viewing time t5 of the keyword course ÷ total duration t6 of the keyword course; If the ratio C is ≥ 90%, the target user's knowledge mastery is judged, and course push adjustments are made based on the target user's knowledge mastery. If the ratio C is less than 90%, the difficulty level of the course will be adjusted.

7. The metadata expansion method based on semantic analysis according to claim 6, characterized in that: The determination of the target user's knowledge mastery and adjustment of course push based on the target user's knowledge mastery include: Obtain the pause, fast-forward, and rewind status of target users during the course viewing process; If there is a pause or fast-forward, the course difficulty level will be adjusted; If there is no pause or fast forward, but there is rewind, the knowledge point push adjustment will be made; If there is no pause, fast forward, or rewind, no adjustment will be made.

8. The metadata expansion method based on semantic analysis according to claim 7, characterized in that: The difficulty level of the courses will be adjusted, including: Classify keyword courses according to the difficulty level of the courses, where the keyword course classification includes difficult and easy; Get the category of the keyword course that the target user is currently watching; If the category of the keyword course currently being viewed is easy, adjust the category of the keyword course to difficult; If the classification of the keyword course currently being watched is difficult, a questionnaire will be pushed to the target user, asking whether the keyword course currently being watched by the target user is easy; If the target user answers simply, then ask the target user whether they understand the knowledge points involved in the keyword course they are currently watching; If the target user answers that they understand, no adjustment will be made; If the target user answers that they do not understand, all knowledge points in the keyword course currently watched by the target user are obtained, and courses corresponding to other knowledge points under the major that do not include all knowledge points in the keyword course currently watched by the target user are pushed to the target user; If the target users have difficulty answering, adjust the classification of the keyword course to easy.

9. The metadata expansion method based on semantic analysis according to claim 8, characterized in that: The knowledge point push adjustment includes: Get the time point before and after rewinding the target user's current course, recorded as P1 and P2 respectively; Extract the knowledge points between time point P2 and time point P1; Refine and organize the extracted knowledge points to form knowledge point keywords; Obtain the courses that correspond to the search keywords, knowledge point keywords, and recommended majors in the target database and define them as knowledge point courses; Get the duration of the knowledge point segment corresponding to the knowledge point keyword in the course currently watched by the target user, recorded as t7; Extract all knowledge points in the target user's current course; Get the duration of each knowledge point segment corresponding to the target user's current course, and record them as M1, M2, ..., Mn respectively; Calculate the sum of the duration of the corresponding knowledge point segments of all knowledge points in the target user's current course, denoted as M, M = M1 + M2 + ... + Mn; Calculate the ratio of the duration t7 of the knowledge point segment corresponding to the knowledge point keyword in the target user's current course to the sum of the durations M of the knowledge point segments corresponding to all knowledge points in the target user's current course, denoted as D. Ratio D = duration t7 ÷ sum of durations M; According to the total number of course pushes N and the ratio D, calculate the number of knowledge point course pushes, recorded as N1, the number of knowledge point course pushes N1 = total number of pushes N × ratio D; According to the total number of course pushes N and the number of knowledge point course pushes N1, calculate the number of keyword course pushes, recorded as N2, the number of keyword course pushes N2 = the total number of pushes N - the number of knowledge point course pushes N1; Adjust the course push according to the number of knowledge point courses pushed N1 and the number of keyword courses pushed N2, and push knowledge point courses and keyword courses to target users.

Images